One vendor in five publishes no headquarters, and one in four no usable rate

Every listing in the index carries a category, a business type, and a live website. Past those three fields the record thins fast: 699 listings name no state and 846 publish no hourly band.

The Wall's August 9 snapshot holds 3,460 approved listings. Three fields on those records come back complete. Every listing carries a category, every listing carries a business type, and every listing carries a live website. Past those three the record thins out fast. 699 listings publish no headquarters state, and 846 publish no hourly rate a buyer can act on.

Field completeness across 3,460 listings

Field	Listings with a value	Share of index
Category	3,460	100%
Business type	3,460	100%
Website	3,461 records	100%
Hourly-rate record present	2,912	84.2%
Headquarters state	2,761	79.8%
Hourly-rate band published	2,614	75.5%

The two gaps are different in kind. The 846 listings without a usable rate split into two groups: 298 carry a rate record marked undisclosed, which is a decision to say nothing, and 548 carry no rate record at all. The headquarters gap has no such split. A listing either names a state or it does not, and 699 do not. That single number is why every geographic briefing published here reports state counts as a floor rather than a full census, including the Census-region rollup.

What follows is interpretation rather than data. The fields that come back complete are the ones the index establishes on its own: the category assignment, the business type, and the live-site check that gates publication in the first place. The fields that come back thin are the ones a vendor has to volunteer. Nothing in the snapshot records why a firm withholds an address or a rate, so the pattern is worth naming and not worth explaining. How the gate itself works is covered in the verification briefing.

For a buyer the cost is concrete. Filter a shortlist to a single state and 699 firms drop out before the first call, not because they sit somewhere else but because they never said. Filter on a published rate band and 846 drop out on the same terms. Both filters look free. Both cut a fifth to a quarter of the market on a blank field rather than a bad answer. The cheap correction is to run the filter, run it again without, and look at what only the second pass returns. Rate-disclosure patterns broken out by pillar sit in the transparency map.

One sourcing note. The figures above are field counts from the index snapshot captured August 9, 2026, not estimates. The snapshot holds 3,461 domain records against a 3,460-listing total, a one-record variance carried through from the capture.